Trains.com

Crewless Railroading on the Horizon

16007 views
75 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    December 2001
  • From: NW Wisconsin
  • 3,857 posts
Posted by beaulieu on Friday, February 24, 2012 9:29 PM

BaltACD

Au Contrair - No automated train will EVER be run without the full tonnage it is scheduled - it would not fit the efficiency model.  You are overlooking that railroads, real, imaginary or automated do have a real finite capacity in the number of trains that can be operated.  Bean counters may not believe it, but they only have to work with numbers - not reality.

 


No track capacity is determined first by train braking ability. However it usually is further degraded by a poor match between signal block length with train length, limitations of the signalling system, Sluggish trains (very low hp/weight ratios), poor crew utilization, etc.  NS trialling of GE's LEADER system suggests that at least they are trying to balance fuel costs against the value of track capacity. Future integration of software like LEADER with CAD (Computer Aided Dispatching, and PTC, should allow better utilization of track capacity. Maybe someday Precision Scheduling might become more than just a slogan for public disinformation.

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Thursday, February 23, 2012 8:06 PM

So what is the point of it?  The article offers no explanation of that seemingly obvious question.  The Rio Tinto position seems rather ambiguous about the issue of saving labor.  They are removing the labor needed to run the trains, and yet they say they are not cutting labor because the operation will grow elsewhere and that will require new labor.  That strikes me as fuzzy math.  Even if new operations require new labor, there would be still more labor in total if they did not remove labor from running the trains.  I get the impression they feel that it would be insensitive to say they are automating trains to eliminate jobs.

 

If a train operation had a lot of variability, I could see where removing the crew might save money by eliminating complications in coordinating operations with crews.  But with a single purpose operation like theirs, running almost like a conveyor belt, it seems to me that they would have such a high level of consistency in train operation and scheduling that it is hard to see the advantage of running without an engineer other than saving the wages.   

  • Member since
    June 2003
  • From: South Central,Ks
  • 7,170 posts
Posted by samfp1943 on Thursday, February 23, 2012 7:24 PM

Seems that Rio Tinto has announced their system implementation " Auto Haul"

TRAINASNewswire of this date carries the article:

"Automated train operations to begin on Australian ore-hauling railroad"

Published: February 23, 2012
FTA:"...PERTH, Australia – Rio Tinto Iron Ore, owner of Australia’s Pilbara Rail, announced it would spend $518 million to introduce unmanned train operations across its main line. Pilbara Rail controls both the Hamersley Iron and Robe River Iron railway operations. The company will use crewless trains to carry iron ore from mines in the Pilbara region of northwest Australia to the Indian Ocean ports of Dampier and Cape Lambert.

Dubbed AutoHaul, the crewless system will be the first long-distance, heavy-haul automated system of its kind in the world. The largest privately owned heavy freight railway in Australia, spanning 869 miles, the Pilbara iron ore operations have an annual capacity of 225 million tons. Plans are to increase to increase to 283 million tons by 2013, and 353 million tons a year in 2015.
 
Successful trials of the AutoHaul technology were conducted in 2007, before being suspended with the onset of the global financial crisis. The project will now be implemented in two stages, with the first crewless, or “autonomous,” trains to begin operation in late 2014.

The company said that with its present expansion plans, the railroad staff would require more people, not fewer, to help the business grow. It is not planning any reductions in staff as a result of the autonomous train process.
 
Pilbara Rail has been working with the Australia Office of Rail Safety to ensure the safety of the system. The company said, “We are confident that with the operation of road crossing gates, cameras, motion sensors and other similar initiatives, that AutoHaul will achieve the highest levels of safety.”

The following statement posted earlier by M636C  pretty much sums up what I was thinking about:

"...Exactly what is a driver expected to with a train with 25000 tons of iron ore approaching a grade crossing at 50 mph if a vehicle tries to cross too close in front of him? Not even ECP brakes will do much there (although the couplers probably wouldn't break).

The adjacent  BHP Billiton line had some notices at grade crossings. They stated "Our trains take five minutes to pass through this crossing, whether or not your vehicle is on it at the time...."

M636C


The latter statement possibly is prophetic of future events, I know nothing about Australian Law, but I would suspect that their 'Legal Eagles' will be on alert for bad outcomes at grade crossings(?)My 2 Cents

 

 


 

  • Member since
    September 2003
  • From: Alexandria, VA
  • 847 posts
Posted by StillGrande on Thursday, February 23, 2012 2:22 PM

I think it is a valid point raised that crewless trains would eliminate trains running out of hours on the crew and the subsequent delays in getting a new crew to the train.  If anything yard capacity would have to be further increased because the trains would not be delayed as often due to human considerations, meaning trains might get to their destinations faster, putting more stress on yard functions.

I would also think that the railroads would consider contracting out their train running repairs.  They might start stashing a knuckle every 5-10 cars or so, or on a particular type of car, so that the parts are available for use (I am not sure why this is not done now except for the expense, but when you consider the need and the lack of crew at the head end, then it becomes more cost effective).  A repair service would bid and contract for a set territory.  The contractor would have to meet repair times, worry about getting their people to and from the job site (which would probably be pinpointed by the sensors on the train itself, probably including the problem and a suggestion about what needs to be fixed), and all the associated costs of keeping a set of employees ready to go 24/7, including insurance and retirement.  The railroads would be free of that cost too. 

One would think that the train would still carry event recorders, and should the sensors on the train detect an impact they would still stop and automatically download video to a central location for immediate evaluation.  Trains are going to hit things with or without a crew. 

As far as a computer failure, isn't is common practice now for signals and brakes to fail to the safe position?  Why would the train not fail to stop?  That is what it does now.

I know computers don't wave back.  Maybe that is the problem.

I can't foresee local switch jobs being automated.  Having them automatically couple and uncouple could happen though.  That might just eliminate another person from train service.

Dewey "Facts are meaningless; you can use facts to prove anything that is even remotely true! Facts, schmacks!" - Homer Simpson "The problem is there are so many stupid people and nothing eats them."
  • Member since
    March 2003
  • From: Central Iowa
  • 6,901 posts
Posted by jeffhergert on Thursday, February 23, 2012 10:56 AM

There's only so many engines available at most terminals at any one time.  Shorting a train to wait for more tonnage at one location could cause delays at the next location because they don't have the available engines to move the tonnage they have ready. 

It may not be as much of a problem at the larger terminals that originate/terminate many trains a day.  It's those that only originate a few trains per day could be caught short.

Jeff

  • Member since
    December 2001
  • 8,156 posts
Posted by henry6 on Thursday, February 23, 2012 8:51 AM

beaulieu

Automated trains mean that there would be no reason to hold trains for tonnage, no running out of crews at a terminal, easier to run trains to a schedule since a manifest could be dispatched on time with the cars that were ready and a portion of the locomotives that would be allocated. Say the manifest was scheduled for three locomotives and as departure time approached one third of the cars weren't ready, you would dispatch the two-thirds that are ready with two of the locomotives, and send the remaining third with the final locomotive when its ready.

It is doubtful your assessment has any validity.  Crewless trains are more a function of on the road operation rather than pre departure make up and availablity.  I would think that with automation the number of trains and scheduling would determine its efficiency and that afterthoughs would mess up the system in the long run.

RIDEWITHMEHENRY is the name for our almost monthly day of riding trains and transit in either the NYCity or Philadelphia areas including all commuter lines, Amtrak, subways, light rail and trolleys, bus and ferries when warranted. No fees, just let us know you want to join the ride and pay your fares. Ask to be on our email list or find us on FB as RIDEWITHMEHENRY (all caps) to get descriptions of each outing.

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Thursday, February 23, 2012 6:54 AM

Au Contrair - No automated train will EVER be run without the full tonnage it is scheduled - it would not fit the efficiency model.  You are overlooking that railroads, real, imaginary or automated do have a real finite capacity in the number of trains that can be operated.  Bean counters may not believe it, but they only have to work with numbers - not reality.

beaulieu

Automated trains mean that there would be no reason to hold trains for tonnage, no running out of crews at a terminal, easier to run trains to a schedule since a manifest could be dispatched on time with the cars that were ready and a portion of the locomotives that would be allocated. Say the manifest was scheduled for three locomotives and as departure time approached one third of the cars weren't ready, you would dispatch the two-thirds that are ready with two of the locomotives, and send the remaining third with the final locomotive when its ready.

Never too old to have a happy childhood!

              

  • Member since
    March 2003
  • From: Central Iowa
  • 6,901 posts
Posted by jeffhergert on Thursday, February 23, 2012 6:27 AM

henry6

 Bucyrus:

 

The reasons Rio Tinto is said to be going driverless are shortening journey times by eliminating stops for crew changes, and reducing energy consumption and CO2 emissions. They don’t say anything about eliminating crew cost. 

 

Do they have to say so?  Or since it is a one man crew, he will actually be "operating" the train with a joy stick from a safe, warm, non moving, all amentites room in the Carribean?  Or someplace similar.

If railroads are going to have a human actually operating the train, they are going to have him/her on board the train.  It's not that they couldn't operate a train by R/C, but if you are going to have a person operate the train he might as well be on the train to trouble shoot and repair minor problems.  You don't really gain anything by moving the engineer from the cab to India and it may actually cost more to do so on a wide spread basis.  No, if they are crewless, they will be automatic, run by computers between terminal points.  Any people at the master computer at a central location would only be overseeing the entire operation, making sure everything was working and notifying field personnel of any problems to be checked.  Terminal switching would be done by people on the ground, either conventionally or by RCO.  

The thing that's going to make it unattractive to have completely crewless trains in the US is the wait time for a utility man to reach a disabled train and fix the problem.  Doesn't matter if it's next to the highway or they have to use ATVs to drive over hill and dale.  At times one U-man will be too many, others when 5 are not enough on each territory. (The railroad bean counters will only see the times when one is too many to figure their staffing levels.)  With the delay cost to the railroad in the 5 figures per hour range, getting the trains moving ASAP will be imperative. You can't wait 5 hours because the U-man is checking out 3 other trains that have broke down. 

Oh, we talk about burst air hoses and broken knuckles but often the problems needing attention are mechanical.  Some things can be reset, which could be done by remote.  Others actually require the services of person.  This happens more often than one would think.  If you are superstitious, you would say these things happen in groups because it seems either no one is having problems or everyone is having problems.

Jeff       

  • Member since
    December 2001
  • From: NW Wisconsin
  • 3,857 posts
Posted by beaulieu on Wednesday, February 22, 2012 9:46 PM

Automated trains mean that there would be no reason to hold trains for tonnage, no running out of crews at a terminal, easier to run trains to a schedule since a manifest could be dispatched on time with the cars that were ready and a portion of the locomotives that would be allocated. Say the manifest was scheduled for three locomotives and as departure time approached one third of the cars weren't ready, you would dispatch the two-thirds that are ready with two of the locomotives, and send the remaining third with the final locomotive when its ready.

  • Member since
    November 2005
  • 4,190 posts
Posted by wanswheel on Wednesday, February 22, 2012 9:04 PM

Excerpt from The Automated Times Square-Grand Central Shuttle by Mark Feinman

The automated train was actually running during the fall of 1961 without passengers, running tests to ensure its safety and operation. On January 4th, 1962 at 3:17 pm lasting until about 7 pm, the first automated train in the US began revenue service. Although not needed, in deference to the TWU, a motorman rode the train at all times but did not operate it. The following day the train ran according to the regular shuttle timetable for track 4. Signs were posted all over track 4 indicating that the train was being automatically operated. The motorman did not ride in his cab. The public didn't seem to care (except, of course, for the railfans).

http://www.nycsubway.org/lines/irtshuttle.html

  • Member since
    December 2001
  • 8,156 posts
Posted by henry6 on Wednesday, February 22, 2012 7:07 PM

Bucyrus

 

The reasons Rio Tinto is said to be going driverless are shortening journey times by eliminating stops for crew changes, and reducing energy consumption and CO2 emissions. They don’t say anything about eliminating crew cost. 

Do they have to say so?  Or since it is a one man crew, he will actually be "operating" the train with a joy stick from a safe, warm, non moving, all amentites room in the Carribean?  Or someplace similar.

RIDEWITHMEHENRY is the name for our almost monthly day of riding trains and transit in either the NYCity or Philadelphia areas including all commuter lines, Amtrak, subways, light rail and trolleys, bus and ferries when warranted. No fees, just let us know you want to join the ride and pay your fares. Ask to be on our email list or find us on FB as RIDEWITHMEHENRY (all caps) to get descriptions of each outing.

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Wednesday, February 22, 2012 6:55 PM

henry6
If automation can be used to make money, then yes, do it, as long as it improves the product quality or safety of delivery. 

The reasons Rio Tinto is said to be going driverless are shortening journey times by eliminating stops for crew changes, and reducing energy consumption and CO2 emissions. They don’t say anything about eliminating crew cost. 

  • Member since
    December 2001
  • 8,156 posts
Posted by henry6 on Wednesday, February 22, 2012 6:21 PM

Helicopters already get where there are no roads.  They can drop men and equipment and they can either ride with the rerailed or repaired train or wait for a pick up.  Then there are quite a few prototype jet packs which can carry a man in; and by remote control, it could become a drone carrying things needed.  It is amazing where we are and how close we are to so much more.  If automation can be used to make money, then yes, do it, as long as it improves the product quality or safety of delivery.  But there has to be a point where it might not be worth it and that point has to be defined and identified.

RIDEWITHMEHENRY is the name for our almost monthly day of riding trains and transit in either the NYCity or Philadelphia areas including all commuter lines, Amtrak, subways, light rail and trolleys, bus and ferries when warranted. No fees, just let us know you want to join the ride and pay your fares. Ask to be on our email list or find us on FB as RIDEWITHMEHENRY (all caps) to get descriptions of each outing.

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Wednesday, February 22, 2012 5:19 PM

n012944

 Bucyrus:

There will be no more carrying knuckles through the snow.  The mobil repair service will be like a traveling pit stop that goes where it is needed.       

 

 

Really?  So what part of this new tech will make broken knuckles only happen at grade crossings?  There are plenty of places where the only way the right of way can be gotten to is on the rail.  

I think they can mechanize the process to the point where they can get power equipment to the site of the break no matter where it is.  They might not get a truck to the spot, but they could develop power equipment that would be an extension of what the truck can reach. 

I don't think this will happen overnight.  And it won't happen in a vacuum.  It will just be a gradual part of a much larger process of development.  And it won't just happen as a universal changeover.  And because it is a labor issue, there will be tremendous resistance, aside from the technical hurdles.

But they are working on coupling and uncoupling freight cars by remote control in order to reduce danger.  That seems like a rather far fetched objective with lots of technical issues.  If they are serious about that, remote running doesn't seem like that big of a reach.  There are people working on developing almost everything you can imagine.  I am not advocating remote or automatic running.  I am not even sure what the cost benefit would be.  But it does seem to be a idea that continues to germinate.

And every big change that can be seen coming is always met by disbelief that it is possible.    

  • Member since
    August 2004
  • From: The 17th hole at TPC
  • 2,283 posts
Posted by n012944 on Wednesday, February 22, 2012 4:49 PM

Bucyrus

There will be no more carrying knuckles through the snow.  The mobil repair service will be like a traveling pit stop that goes where it is needed.       

 

Really?  So what part of this new tech will make broken knuckles only happen at grade crossings?  There are plenty of places where the only way the right of way can be gotten to is on the rail.  

An "expensive model collector"

  • Member since
    July 2006
  • 9,610 posts
Posted by schlimm on Wednesday, February 22, 2012 4:45 PM

The tipping point will be when on major routes track capacities reach limits and automated systems advance to a point where that capacity could be increased considerably with its implementation.

C&NW, CA&E, MILW, CGW and IC fan

  • Member since
    February 2003
  • From: Guelph, Ontario
  • 4,819 posts
Posted by Ulrich on Wednesday, February 22, 2012 4:23 PM

I think so...from what I've read PTC borders on the impossible by the mandated deadline, and crewless trains would be several orders of magnitude more complex and expensive... I can see it for simple remote operations, but not for awhile in heavily congested corridors.

  • Member since
    January 2002
  • 4,612 posts
Posted by M636C on Wednesday, February 22, 2012 4:19 PM

oltmannd

 

 tree68:

 

 

Of course, this requires railroading in a fairly sterile environment, as has already been discussed.  Adding in all the other variables we know exist (crossings, pedestrians, etc) is another issue entirely, but a significant one.

The horizon is a long, long way off...

 

 

 

The train handling part is definitely doable now - has been for quite a while.  The part that isn't is the "real life" stuff.  What happens when real world intrudes on the railroad is the trick. If all you had to do was go from A to B and whistle for the road Xings, we could do that right now.  But everything from sweeping out switch points in the snow to reacting to a road Xing... that's the rub.

I can see getting to quite a few one man crews with PTC-based safety controls as a back up, but crewless is quite a bit more than that.

 

Exactly what is a driver expected to with a train with 25000 tons of iron ore approaching a grade crossing at 50 mph if a vehicle tries to cross too close in front of him? Not even ECP brakes will do much there (although the couplers probably wouldn't break).

The adjacent  BHP Billiton line had some notices at grade crossings. They stated "Our trains take five minutes to pass through this crossing, whether or not your vehicle is on it at the time."

M636C

  • Member since
    December 2001
  • 8,156 posts
Posted by henry6 on Wednesday, February 22, 2012 4:11 PM

I would thnk crewless trains on the whole would require inch by inch east/west (seeing in both directions as well as peripheral views) visual monitoring along with sattellite GPS and visual.  WHen does it cost too much, where is the break point, the point of no return?  What an engineering nightmare!  What a progamming nightmare!  This system will buy a lot of health insurance, pay for a lot of cars, houses and refrigerators, and buy a lot of victuals perhaps with more value than getting rid of a person on the payroll!

 

RIDEWITHMEHENRY is the name for our almost monthly day of riding trains and transit in either the NYCity or Philadelphia areas including all commuter lines, Amtrak, subways, light rail and trolleys, bus and ferries when warranted. No fees, just let us know you want to join the ride and pay your fares. Ask to be on our email list or find us on FB as RIDEWITHMEHENRY (all caps) to get descriptions of each outing.

  • Member since
    February 2003
  • From: Guelph, Ontario
  • 4,819 posts
Posted by Ulrich on Wednesday, February 22, 2012 4:04 PM

Any project that involves  more than two people tends to get balled up in hurry as well.

  • Member since
    January 2008
  • 1,243 posts
Posted by Sunnyland on Wednesday, February 22, 2012 3:04 PM

I don't  like this idea.  Anything computerized can malfunction and then what happens. Runaway trains?

Even remote control locos are a concern to me.  Sometimes things happen that only a human on site can take care of, not a computer or someone sitting in front of a computer terminal.

And there have been enough job losses in this country with outsourcing and now they want to try this. Just so the "big bosses" can put more money in their pockets and the heck with the lowly working man/woman.

  • Member since
    January 2001
  • From: Atlanta
  • 11,971 posts
Posted by oltmannd on Wednesday, February 22, 2012 11:40 AM

tree68

 

 samfp1943:
Maybe, Zug, that is your calling....To develop a 'dyno/sensor feed'  that can measure the tractive effort of the Engineer's relation to his seat cushion as a method to determine train handling applications(?) Mischief

 

 

 

While I'm not looking to put anyone out of of job, I would submit that the computing power exists today to allow for an autonomous train.  What's lacking is the data necessary to do so. 

Such an operation would require precise data on the track - grades, curves, etc.  This would be fairly easy to acquire - a trip over the territory in question with the appropriate sensors would provide all the information needed.

Even the "seat of the pants" sensor could be had.

What's needed is a way to accurately measure the parameters of each individual car - weight, slack dynamics, braking capabilities, you name it.  Given the number of cars in circulation, this is an almost insurmountable barrier.  One that makes continuing to put human crews on board the trains more appealing.

Given the appropriate data, the motive power (as well as the individual cars) could be commanded  to take the appropriate action at the appropriate time.

Taking electronic braking to the next level, cars could be commanded to apply or release their brakes individually as needed/appropriate.  This could serve to minimize slack run-ins, especially in "hogback" profiles.

Of course, this requires railroading in a fairly sterile environment, as has already been discussed.  Adding in all the other variables we know exist (crossings, pedestrians, etc) is another issue entirely, but a significant one.

The horizon is a long, long way off...

 

The train handling part is definitely doable now - has been for quite a while.  The part that isn't is the "real life" stuff.  What happens when real world intrudes on the railroad is the trick. If all you had to do was go from A to B and whistle for the road Xings, we could do that right now.  But everything from sweeping out switch points in the snow to reacting to a road Xing... that's the rub.

I can see getting to quite a few one man crews with PTC-based safety controls as a back up, but crewless is quite a bit more than that.

-Don (Random stuff, mostly about trains - what else? http://blerfblog.blogspot.com/

  • Member since
    December 2001
  • From: Northern New York
  • 25,023 posts
Posted by tree68 on Wednesday, February 22, 2012 9:59 AM

samfp1943
Maybe, Zug, that is your calling....To develop a 'dyno/sensor feed'  that can measure the tractive effort of the Engineer's relation to his seat cushion as a method to determine train handling applications(?) Mischief

While I'm not looking to put anyone out of of job, I would submit that the computing power exists today to allow for an autonomous train.  What's lacking is the data necessary to do so. 

Such an operation would require precise data on the track - grades, curves, etc.  This would be fairly easy to acquire - a trip over the territory in question with the appropriate sensors would provide all the information needed.

Even the "seat of the pants" sensor could be had.

What's needed is a way to accurately measure the parameters of each individual car - weight, slack dynamics, braking capabilities, you name it.  Given the number of cars in circulation, this is an almost insurmountable barrier.  One that makes continuing to put human crews on board the trains more appealing.

Given the appropriate data, the motive power (as well as the individual cars) could be commanded  to take the appropriate action at the appropriate time.

Taking electronic braking to the next level, cars could be commanded to apply or release their brakes individually as needed/appropriate.  This could serve to minimize slack run-ins, especially in "hogback" profiles.

Of course, this requires railroading in a fairly sterile environment, as has already been discussed.  Adding in all the other variables we know exist (crossings, pedestrians, etc) is another issue entirely, but a significant one.

The horizon is a long, long way off...

 

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    June 2003
  • From: South Central,Ks
  • 7,170 posts
Posted by samfp1943 on Wednesday, February 22, 2012 9:14 AM

zugmann

 

 AndrewDickey:

 

Keep in mind that the engineer is not controlling the train right now - he (or she) is telling a computer what they want and the computer controls the train.  And as others have said, sometimes the engineer is simply doing what some other computer tells him (or her) is the best thing to do.  The technology is already there.

 

 

 

 

"...It's a very efficient way of ripping trains into multiple pieces.  There has yet to be a computer program that can calculate as well as an engineer's buttcheeks..."

Maybe, Zug, that is your calling....To develop a 'dyno/sensor feed'  that can measure the tractive effort of the Engineer's relation to his seat cushion as a method to determine train handling applications(?) Mischief

 

 


 

  • Member since
    June 2003
  • From: South Central,Ks
  • 7,170 posts
Posted by samfp1943 on Wednesday, February 22, 2012 9:13 AM

zugmann

 

 AndrewDickey:

 

Keep in mind that the engineer is not controlling the train right now - he (or she) is telling a computer what they want and the computer controls the train.  And as others have said, sometimes the engineer is simply doing what some other computer tells him (or her) is the best thing to do.  The technology is already there.

 

 

 

 

"...It's a very efficient way of ripping trains into multiple pieces.  There has yet to be a computer program that can calculate as well as an engineer's buttcheeks..."

Maybe, Zug, that is your calling....To develop a 'dyno/sensor feed'  that can measure the tractive effort of the Engineer's relation to his seat cushion as a method to determine train handling applications(?) Mischief

 

 


 

  • Member since
    January 2002
  • 4,612 posts
Posted by M636C on Wednesday, February 22, 2012 5:07 AM

zugmann

 

It's a very efficient way of ripping trains into multiple pieces.  There has yet to be a computer program that can calculate as well as an engineer's buttcheeks.

In 1978 I was riding a Hamersley Iron loaded train that broke in five places. The train was approaching Seven Mile yard when the yard tower decided to change our allocated track, in the process reversing the approach signal from green to red. The driver made an emergency application and the train broke in five places. Four of the breaks were coupler knuckles but the fifth was a rigid drawbar between a married pair of cars. The drawbar was fine but the draft gear yoke had failed. But we stopped short of the signal.

That was 220 100 ton cars hauled by a C36-7 and two M636 units.

We had instruments on the 199th and 200th cars with coaxial cable rigged down the train, held on by five cable ties on each car.  That was two kilometres of cable at one dollar per metre in those days. We never saw any of it again.

Of course the operation is called Rio Tinto now...

But they have been investigating the effects of train handling for nearly forty years to my own personal knowledge. Probably hundreds of trains have been recorded, measuring coupler tension and compression forces, train speeds at different locations, truck angles in curves locomotive power setting and many other details.

They no longer use fixed signals. There is an in cab signalling system that tells the driver where he is, what the maximum allowed speed is and what is the recommended speed for existing conditions.

Combining this with a computer control system based on the decades of accumulated train handling information, and the ability to match the control inputs to the exact location on the track, I would be very surprised if a train was to break now.

There have been experimental systems running on the line since 2006 or earlier. They were close to introducing the system in 2008 when they decided to cancel the project due to the Global Financial Crisis and the consequent downturn (not that that had any effect on Rio Tinto iron ore production, which just kept increasing owing to demand in China).

The system they are using is quite elegant. The control system is contained in a box that fits into the walkway at the front of the lead unit and it is connected to the locomotive through the multiple unit cables. So the locomotive reacts as though connected to a unit ahead with a normal crew. No permanent change to the locomotive at all.

There is a distinctive antenna mast mounted against the locomotive nose on the left side.

Interestingly, I have seen a photo of exactly that system mounted on a Union Pacific unit at GE's test track in Erie Pennsylvania in 2007. So this isn't a strange foreign experiment. It is an American system, tested in the USA. And it can be fitted easily to any standard locomotive....

While the Hamersley line runs through sparsely settled areas, it has serious grades, a rising 0.42% against loaded trains, but a ten mile long 2% falling grade where good train control is important.

Rio Tinto don't use distributed power. It is the parallel line to the East, BHP Billiton that does that. The trains on Rio Tinto generally have two locos at the head and about 240 cars of 100 ton capacity.

M636C

  • Member since
    January 2002
  • From: Canterlot
  • 9,575 posts
Posted by zugmann on Wednesday, February 22, 2012 1:16 AM

AndrewDickey

Keep in mind that the engineer is not controlling the train right now - he (or she) is telling a computer what they want and the computer controls the train.  And as others have said, sometimes the engineer is simply doing what some other computer tells him (or her) is the best thing to do.  The technology is already there.

 

 

It's a very efficient way of ripping trains into multiple pieces.  There has yet to be a computer program that can calculate as well as an engineer's buttcheeks.

It's been fun.  But it isn't much fun anymore.   Signing off for now. 


  

The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.t fun any

  • Member since
    January 2002
  • From: Canterlot
  • 9,575 posts
Posted by zugmann on Wednesday, February 22, 2012 1:12 AM

Well, look at it this way.  By the time we completely automate the railroads and trucking industry, we won't need them since no one in this country will be working (except a few computer people), so there will be no need to move large amount of goods. 

 

 

It's been fun.  But it isn't much fun anymore.   Signing off for now. 


  

The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.t fun any

  • Member since
    March 2003
  • From: Central Iowa
  • 6,901 posts
Posted by jeffhergert on Tuesday, February 21, 2012 11:12 PM

AndrewDickey

Keep in mind that the engineer is not controlling the train right now - he (or she) is telling a computer what they want and the computer controls the train.  

 

Really?  Could you please elaborate on this, and how it works.

Jeff 

Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy